JP3343104B2 - Image processing apparatus and image processing method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing apparatus and method for compressing / decompressing image data, and more particularly to a printer capable of gradation-recording multi-valued image data having continuous gradation read by a scanner or the like. The present invention relates to an image processing apparatus that can be applied to a copying machine that uses the same.

[0002]

2. Description of the Related Art In a conventional digital copying machine, image data read from a scanner unit is stored in an image memory, and after editing such as rotation or synthesis is read out from the image memory and printed out from a printer unit. For example, when recording paper is fed vertically by the printer against a document placed horizontally on the reading surface of the scanner, the image data read from the scanner is rotated to change the orientation of the recording paper. A flexible response such as matching was possible.

On the other hand, a multi-value printer capable of recording not only binary data but also intermediate data has been developed. Multi-valued printers express image data that has been represented by one bit per pixel in the past as multi-valued data of two bits or more per pixel, so that continuous tone images can be recorded and image quality can be improved. it can.

[0004] When handling multi-valued image data, the data amount is several times larger than that of binary image data, so that the storage capacity of an image memory for storing image data is increased, resulting in an increase in cost. If the image data is stored after being compressed, the memory capacity can be reduced. However, depending on the compression method, there is a problem that the compressed image data cannot be edited. For example, JPE, a widely used compression method
Since G encoding is variable-length encoding, information on pixel positions is not stored and editing is difficult. Therefore, as shown in FIG. 15, editing must be performed before the JPEG encoding. However, even with this method, the cost of the editing memory increases because the capacity of the editing memory increases. In addition, there is a disadvantage that the processing speed is reduced unless memory access is speeded up because of a large amount of editing data.

Further, in the case of a digital multifunction peripheral having both a copy function and a facsimile transmission / reception function, if a compression method such as JPEG which is different from the JBIG method, which is one of the compression methods of the facsimile standard, is adopted, two-system encoding is performed. -Since the decoding system must be mounted, there is a problem that the circuit scale is increased and the cost is increased.

To solve such a problem, a method of compressing, accumulating, and restoring multi-level continuous tone data after halftone processing has been considered. As shown in FIG. 16, first, image data is converted into binary data by halftone processing, and the binary data is rotated by block rotation processing using the edit memory 11. After that, facsimile standard coding (JBIG coding) is performed, and the compressed image is stored in the image storage memory 12. Then, the compressed image is read from the image storage memory 12, decoded (JBIG decoding), multi-valued image data is estimated from the restored halftone data by multi-value estimation processing, and output as multi-valued continuous tone data. .

[0007]

However, as described above, the method of compressing / accumulating / decompressing after halftone processing is performed by using a spatial filter having a low-pass type spatial frequency characteristic as a method of estimating multi-valued image data. Because of the use, the restored image is blurred, but the low-frequency unevenness generated in the halftone process cannot be sufficiently attenuated, and there is a problem that the image quality is greatly deteriorated.

SUMMARY OF THE INVENTION The present invention has been made in view of the above circumstances, and can be accurately and without image quality deterioration even after editing after image compression, and can suppress an increase in required memory capacity. It is an object of the present invention to provide an image processing apparatus that realizes image encoding / decoding that is compatible with encoding / decoding of facsimile standards such as JBIG.

[0009]

According to the present invention, the orthogonal transform coefficient is converted into a predetermined number of bits for each spatial frequency, and the quantized data is rearranged into a banded arrangement for each spatial frequency, so that the same data is obtained between adjacent blocks. Bit serial data in which quantized data in the spatial frequency band is continuous is generated, and the bit serial data is compressed.

This makes it possible to suppress an increase in the required memory capacity by editing the quantized data. In addition, bit serial data in which quantized data of the same spatial frequency band continues between adjacent blocks is defined by a facsimile standard such as JBIG. Also suitable for encoding / decoding.

[0011]

DESCRIPTION OF THE PREFERRED EMBODIMENTS A first aspect of the present invention is an orthogonal transform means for decomposing multi-valued image data cut out to a predetermined block size into orthogonal transform coefficients corresponding to spatial frequencies.
Quantizing means for transforming the orthogonal transform coefficients into a predetermined number of bits for each spatial frequency; and quantizing data of the same spatial frequency band between adjacent blocks by rearranging the quantized data into a banded arrangement for each spatial frequency. Is an image processing apparatus comprising: banding means for generating bit serial data which is continuous; and encoding means for compressing the bit serial data.

According to the image processing apparatus configured as described above, the increase in the memory capacity required by editing the quantized data can be suppressed, and the quantized data of the same spatial frequency band is continuously transmitted between adjacent blocks. The bit serial data to be processed has an effect that it is suitable for encoding / decoding of facsimile standards such as JBIG.

According to a second aspect of the present invention, there is provided the image processing apparatus according to the first aspect, further comprising an editing unit for performing image editing on a block-by-block basis on the quantized data calculated by the quantizing unit. Are rearranged into an arrangement in which the quantized data is banded for each spatial frequency by the banding means.

Thus, since the quantized data compressed by the quantization of the orthogonal transform coefficients is edited, an increase in the memory capacity required for the editing memory can be suppressed.

According to a third aspect of the present invention, in the image processing apparatus according to the first or second aspect, a decoding means for expanding the data compressed by the coding means, and a block from the decoded bit serial data Frequency component restoring means for restoring the quantized data of each spatial frequency component for each time; dequantizing means for dequantizing the restored quantized data of each spatial frequency component to restore an orthogonal transform coefficient; Inverse orthogonal transform means for restoring the original block image from the orthogonal transform coefficients,
And block combining means for combining the restored block images to generate original multi-valued image data.

According to such a configuration, in a series of processes such as compression / accumulation / decompression of multi-valued image data, the compressed image data can be edited, and the compressed image data can be decompressed and subjected to block synthesis. Thus, an edited image can be output.

According to a fourth aspect of the present invention, in the image processing apparatus according to the third aspect, the editing means determines a write address or a read address of a memory in which the quantized data is written, by changing a rotation amount and / or a rotation direction of the image. The block synthesizing unit controls the write address or the read address of the memory in which the restored block image is written according to the rotation control data in accordance with the rotation control data. The image in the block is rotated.

Thus, the image data compressed by the quantization of the orthogonal transform coefficients can be rotated and edited, and the edited contents can be operated by an external control signal such as rotation control data.

According to a fifth aspect of the present invention, in the image processing apparatus of the fourth aspect, rotation information indicating the content of rotation control is added to the bit serial data of the image rotated by the editing means in page units. And an information detecting means for detecting the rotation information from the decoded bit serial data, wherein the block synthesizing means includes an image in the block according to the rotation information detected by the information detecting means. This controls the rotation of.

Thus, the rotation information is added to the bit serial data of each block in page units in the compression process, and the rotation information is detected for each block in the decompression process and used for rotation control. Even if different rotation edits are performed in page units, rotation information can be used in block and page units according to the rotation edits in the compression process by using the rotation information in the restoration process.

According to a sixth aspect of the present invention, in the image processing apparatus according to the third, fourth or fifth aspect, the coding method of the coding means and the decoding method of the decoding means are facsimile coding / decoding. System.

[0022] With this configuration, the data is converted into bit serial data in a format suitable for JBIG encoding, which is one of the facsimile standards, at the preceding stage of the encoding means. Even if the decoding method of the decoding means is a facsimile encoding / decoding method, further compression becomes possible and decompression can be performed by the facsimile encoding / decoding method.

According to a seventh aspect of the present invention, in the image processing apparatus according to the third to sixth aspects, there is provided a halftone processing means for converting the multi-valued image data into halftone data, and whether a copy operation or a facsimile transmission is performed. Function selecting means for switching data to be input to the encoding means from the halftone data and the bit serial data in accordance with a designated function selection signal, wherein the facsimile transmission is selected when facsimile transmission is selected. Is externally output.

Thus, when the encoding means is used as a facsimile encoder, a function can be selected between a copy operation and a facsimile transmission.

According to an eighth aspect of the present invention, there is provided an image input means for scanning a document and inputting image data;
The image processing apparatus of the aspect, a recording unit that records the image restored by the image processing apparatus, and a communication unit that performs facsimile transmission of an image instructed to perform facsimile transmission among images compressed by the image processing apparatus, Is a multifunction peripheral provided with:

According to a ninth aspect of the present invention, the multi-valued image data cut into a predetermined block size is decomposed into orthogonal transform coefficients corresponding to spatial frequencies, and the orthogonal transform coefficients are converted into a predetermined number of bits for each spatial frequency. The quantized data is rearranged into an arrangement in which the quantized data is banded for each spatial frequency to generate bit serial data in which quantized data of the same spatial frequency band continues between adjacent blocks, and the bit serial data is compressed. An encoding method characterized by the following.

According to a tenth aspect of the present invention, the multi-valued image data cut into a predetermined block size is decomposed into orthogonal transform coefficients corresponding to spatial frequencies, and the orthogonal transform coefficients are converted into a predetermined number of bits for each spatial frequency. The quantized data is rearranged into an arrangement in which the quantized data is banded for each spatial frequency to generate bit serial data in which quantized data of the same spatial frequency band continues between adjacent blocks, and the bit serial data is compressed. On the other hand, the compressed data is read and decoded, the quantized data of each spatial frequency component is restored from the decoded bit serial data for each block, and the restored quantized data of each spatial frequency component is restored. Inverse quantization restores the orthogonal transform coefficients, restores the original block image from the restored orthogonal transform coefficients, and synthesizes the restored block image. The image processing method characterized by generating an original multi-valued image data.

Hereinafter, embodiments of the image processing apparatus according to the present invention will be specifically described with reference to the drawings.

(First Embodiment) FIG. 1 is a block diagram relating to a series of processes of compression, accumulation, and decompression of an image processing apparatus according to a first embodiment. In the first embodiment, the JB of the facsimile standard
The IG method is used.

In the image processing apparatus shown in FIG. 1, image data is sequentially written to the first line memory 101. The image data in this example is multi-level data of continuous tone in which one pixel is expressed by 8 bits, but is not limited to 8 bits. The block dividing unit 102 controls the read address of the first line memory 101 and outputs block data of 4 × 4 pixels to the Haar transform unit 103 in block units. The Haar transform unit 103 performs an HAAR transform on the block data of 4 × 4 pixels, which is a kind of orthogonal transform, to obtain an HA.
Get the AR coefficient.

FIG. 2 shows an example of a base pattern used for the HAAR conversion. The HAAR conversion is a process of converting 4 × 4 image data (Dxy) into 4 × 4 HAAR coefficients (HAmn). Input image position (x, y) and HA
The following operation is performed using a base pattern (Pmnxy) determined by the base vector number (m, n) of the AR coefficient.

[0032]

(Equation 1) FIG. 3 is a diagram showing HAAR coefficients obtained by the HAAR conversion. The lower left HAAR coefficient [HA00] is a DC component and is arranged such that the frequency component increases toward the diagonally located HAAR coefficient [HA33]. In this example, each HAAR coefficient (HAmn) is represented by 8 bits. HA calculated by HAAR conversion section 103
The AR coefficient (HAmn) is input to the quantization unit 104.

The quantization unit 104 quantizes the 8-bit HAAR coefficient with the number of quantization bits determined according to the frequency component and the position on the 4 × 4 matrix. FIG. 4 shows a specific example of the quantization process. As shown in the figure, the number of quantization bits is allocated by assigning the number of bits (8 bits in this example) necessary for expressing the continuous tone to the DC component, and the AC component is composed of a lower frequency component than a high frequency component. Many bits are allocated. This is because the AC component is generally small in a halftone image, so that even if the AC component is roughly quantized, the effect on the halftone image quality is small.
This is because although there are many components, high bit precision is not required as in a halftone. Further, in a general image, the number of quantization bits is set to 0 in this example because the oblique component of the high frequency is small and the influence on the image quality is small.

The block data generator 105 receives the quantized data quantized as described above from the quantizer 104. Then, a DC component (QD) 8 bits, an AC low frequency component (QB) 8 bits, and an AC high frequency component (QA) 1
A total of 32 bits of 6 bits are generated as block data (E). The block data (E) is written into the second line memory 107 by the frequency banding unit 106.

As described above, the image data is subjected to Haar transform, the HAAR transform coefficients are quantized and converted into quantized coefficients, whereby the 16 × 8-bit image data per block is converted into 32-bit data (quantized coefficient ). Since the compression processing up to this point is fixed-length encoding, this compressed data can be restored even if editing such as rotation is performed, and the editing memory is also adjusted to the data capacity after compression. It can be reduced to a small capacity.

FIG. 5 is a diagram showing the data arrangement of the quantized data written in the second line memory 107. 1
The quantized data of the block is arranged such that the DC component (QD), the AC low frequency component (QB), and the AC high frequency component (QA) are banded by frequency, and the bit serialized data of the individual frequency components is 2 is written to the second line memory 107. By reading consecutive addresses in the main scanning direction (horizontal direction) in the arrangement diagram shown in FIG. 5, data of the same frequency band can be output as a bit serial string between adjacent blocks.

When reading data of the same frequency band between adjacent blocks from the second line memory 107 as a bit serial string,
The C band part is coded so that 0 of each AC component data corresponds to the white dot of the bitmap, and the DC band part is D code.
When the C component data is in a white state of FF (hex), patterning is performed so as to correspond to eight white dots.

FIG. 6 shows actual bitmap data in which the quantization coefficients of the halftone image are banded by frequency. One square separated by a broken line corresponds to one block. Although a black dot image appears irregularly in the DC band, the occurrence of the black dot image is reduced in the AC low frequency band, and almost all are in the white spot state in the AC high frequency band. This is because a high frequency component is small in a halftone image or a background portion of a character. Also, no AC component is generated in the background portion of the text document,
The white background (255 levels with 8-bit input) is the DC band,
Both AC bands are continuous images of white dots.

In a normal image, white dots frequently occur in a band or block shape. In JBIG coding, the compression efficiency increases as the number of continuous white or black dots increases two-dimensionally. Therefore, as shown in FIG. 6, a binary image pattern in which white dots frequently occur in a band or block shape is a JBIG code. It can be said that this is the image that can increase the compression efficiency the most for image processing.

As described above, according to the present invention, a bitmap pattern capable of performing the editable compression by the first-stage compression process and increasing the compression efficiency most for the JBIG encoding performed in the second-stage compression process is described. Has been generated.

The bit-serialized binary image data is read out from the second line memory 107 and encoded by the JBIG encoding unit 108 according to the JBIG method of the facsimile standard, and the encoded data is stored in the image storage memory 109 in page units. I do. The page data compressed by the JBIG encoding is subjected to page rearrangement and electronic filing on the image storage memory 109.

Next, the compressed page data is read out from the image storage memory 109 and the JBIG decoding unit 110
Is restored to bitmap data by JBIG decoding by Then, the data is written to the third line memory 111 by the block data restoration unit 112. The block data restoration unit 112 controls the write address so as to restore the bitmap pattern arrangement shown in FIG.

When the four lines of restored data are written to the third line memory 111, the frequency component restoring unit 113
Read data of 32 bits per block in units of 8 × 4 bits blocks indicated by solid lines in FIG.
DC component (DD) 8 bits, AC low frequency component (DB) 8
The 16 bit frequency component of the AC high frequency component (DA) is restored. DC component (DD), AC low frequency component (DB), AC
The high frequency component is input to the inverse quantization unit 114 in block units.

FIG. 7 is a conceptual diagram of the inverse quantization performed in the inverse quantization unit 114. As shown in the drawing, the inverse quantization unit 114 restores the HAAR coefficients by converting the quantized coefficients to 8 bits by simple proportional calculation in accordance with the number of quantization bits determined according to the frequency and position conditions as in the quantization. I do. The restored HAAR coefficients are provided to the inverse Haar transform unit 115.

The inverse Haar transform unit 115 generates the restored HA
The AR coefficient (HBmn) is converted into image data by performing an inverse Haar transform using a base pattern. More specifically, this is a process of converting 4 × 4 HAAR coefficients (HBmn) into 4 × 4 image data (Rxy). The base vector number (m, n) of the input HAAR coefficient and the output image data The following calculation is performed by the base pattern (Pmnxy) based on the position (x, y).

[0046]

(Equation 2) Image data (Rxy) obtained by inverse Haar transform
Is written into the fourth line memory 116 by the block synthesizing unit 117 in block units. Then, by reading out the image data from the fourth line memory 116 in line units, the restored image data is output.

As described above, according to the present embodiment, the first-stage data compression is realized by fixed-length encoding by performing Haar transform of image data in blocks and quantizing the HAAR transform coefficients. Therefore, if the image is edited with the compressed data at this point, the capacity of the editing memory can be suppressed. In addition, quantization coefficients (DC components (8 bits), AC low-frequency components (8
Bit) and AC high frequency components (16 bits) are bit-serialized and arranged by frequency (banding), and bitmap data composed of banded bit-serialized data is subjected to JBIG encoding, which is the second stage compression. Therefore, a higher compression ratio can be realized as compared with the case where non-banded bitmap data is compressed.

(Embodiment 2) Next, an image processing apparatus according to Embodiment 2 of the present invention will be described. The image processing apparatus according to the present embodiment is an example in which the image data is made rotatable in units of 90 degrees at the stage when the HAAR coefficients are quantized to generate the block data.

FIG. 8 is a block diagram relating to a series of processes of compression, accumulation, and restoration of the image processing apparatus according to the present embodiment. Note that portions having the same functions as those of the image processing apparatus according to the first embodiment are denoted by the same reference numerals.

In this embodiment, the first line memory 10
The continuous tone multi-valued image data written in 1 is HAAR
The operation is the same as that of the first embodiment up to the point where the block data (E) is generated by performing the conversion and quantizing the HAAR conversion coefficients. In this embodiment, the block data (E) is written into the page memory 202 by the block rotation edit processing unit 201 with one block as one address. This page memory 202 corresponds to an editing memory. The block rotation edit processing unit 201 reads the block data from the page memory 202 at a rotation position address according to the rotation control data indicating the rotation direction and the rotation amount of the image. The rotation control data may include only the rotation direction information if the rotation amount is fixed, or may include only the rotation amount if the rotation direction is fixed. Furthermore, if the rotation amount and the rotation direction are fixed, only the presence or absence of the rotation control may be used.

FIG. 9 is a conceptual diagram of the block rotation edit processing by the block rotation edit processing unit 201. As shown in the figure, although the block data (E) written to the page memory 202 is not rotating, the block data (E) is controlled by controlling the read address (rotational position address).
Are rotated 90 degrees counterclockwise.

As described above, even when image data of 16 × 8 bits is input, the image data (block data (E)) can be edited in a state of being compressed to 32 bits, so that it is an editing memory. In the case of the present example, the capacity of the page memory 202 can be reduced to 1/4 as compared with the conventional method.

The block data (E) edited by the block rotation edit processing unit 201 is converted to a frequency banding unit 10.
8, the data is written to the second line memory 107 in a state of being banded for each frequency as shown in FIG. Then, data of the same frequency band between adjacent blocks is extracted as a bit serial data string and subjected to JBIG encoding.

On the other hand, the encoded data stored in the image storage memory 109 is JBIG-decoded, and the block data is restored by the block data restoring unit 112. FIG. 10 shows a state in which the block data has been restored.
As shown in the figure, the direction in the block remains unrotated. After the frequency component of the restored block data is restored, inverse quantization and inverse HAAR conversion are performed. The restoration process is the same as that in the first embodiment up to this point.

In this embodiment, 4 × 4 pieces of image data (Rxy) restored by the inverse HAAR conversion are written in the fourth line memory 116 by the block synthesizing rotation unit 203 in block units. Block synthesis rotation unit 203
Reads the image in the block by rotating it 90 degrees counterclockwise from the rotation position address according to the rotation control data from the fourth line memory 116 as shown in FIG.
As a result, as shown in FIG. 10, both the entire page and the image in the block are output as images rotated 90 degrees counterclockwise.

As described above, according to the present embodiment, block data (E) obtained by quantizing and compressing a transform coefficient obtained by orthogonally transforming multi-valued image data is rotated and edited.
The capacity of the page memory 202 for rotation editing is the number of processing pixels ×
Since it is 2 bits, the conventional capacity (the number of processed pixels × 8)
Bit), the memory capacity can be reduced to 1/4.

In the compression process, the block arrangement in one page is rotated by controlling the read address of the page memory 202, and in the restoration process, the image in the block is rotated in the same direction as the page rotation direction. Since the image is rotated, the compression time can be reduced.

Although the second embodiment has been described with reference to rotation as an example of editing processing, an editing function other than rotation can be provided. For example, as shown in FIG. 11, image combining processing for combining two images on the page memory 202 can be performed. Also in this case, since the data amount per block is compressed to 1/4, the page memory 202
Capacity can be greatly reduced.

Third Embodiment Next, an image processing apparatus according to a third embodiment of the present invention will be described. This embodiment is an example in which the image processing apparatus according to the second embodiment has an editing function of changing the rotation direction for each page.

FIG. 12 is a block diagram relating to a series of processes of compression, accumulation, and restoration of the image processing apparatus according to the present embodiment. Note that the same reference numerals are given to portions having the same functions as those of the image processing apparatuses according to the first and second embodiments.

In this embodiment, the first line memory 10
The continuous tone multi-valued image data written in 1 is HAAR
The second embodiment is the same as the second embodiment up to the point where the block data (E) is generated by performing the conversion, quantizing the HAAR conversion coefficients, and performing the banding process on the block data by frequency.
In the present embodiment, after the bitmap data (BT) is generated, rotation information is added to the header of the bitmap data (BT) according to the image rotation control data.

The block rotation edit processing unit 201 has a CPU
Since the rotation control data for the rotation editing is given in page units, the block data is rotated and read in the rotation direction specified for each page.

On the other hand, the rotation information header adding unit 301 receives the rotation information indicating the rotation direction added by the block rotation editing processing unit 201 in page units, and adds the rotation information to the header of the bitmap data of the rotated and edited page. Is added.

As a result, the block rotation edit processing section 20
Even if the content of the rotation edit is changed in page units based on the instruction from the CPU in 1, the content of the rotation edit is added to the bitmap data of each page. The direction of rotation can be determined.

On the other hand, when the encoded data stored in the image storage memory 109 is JBIG-decoded, the rotation information header detector 302 detects the rotation information added to the header of the decoded bitmap data. Then, it is provided to the subsequent block synthesis rotation unit 117. The block synthesizing rotation unit 117 performs rotation reading control using rotation information of a corresponding block when performing rotation reading processing of image data from the fourth line memory 116. That is, since the rotation direction of the page to which the block belongs in the compression process can be detected as the rotation direction of the image in the block, the block in the same direction as the rotation direction of the page in the compression process is detected based on the detected rotation direction. Rotate the inside image.

As described above, according to the present embodiment, in the compression process, rotation information indicating the rotation direction in page units is added to the bitmap data, and in the decompression process, the rotation information is extracted from the header of the bitmap data, and Since the rotation direction of the image can be controlled on a page basis, even if the rotation direction differs for each document, the rotation direction can be restored without separately managing the rotation direction for each document.

(Embodiment 4) Next, an image processing apparatus according to Embodiment 4 of the present invention will be described. In the present embodiment, the JB in the image processing apparatus of the first embodiment is used.
This is an example in which the IG encoding unit 108 can be used in both the copy system and the facsimile system described in the first embodiment.

FIG. 13 is a block diagram related to a series of copying (compression / accumulation / decompression) processes and facsimile halftone processing of the image processing apparatus according to the present embodiment. Note that portions having the same functions as those of the image processing apparatus according to the first embodiment are denoted by the same reference numerals.

In this embodiment, the banding unit for each frequency and the J
A selector 401 for selecting a copy system or a facsimile system as a selection function is provided between the selector 401 and the BIG encoding unit 108. The selector 401 includes the halftone processing unit 4
02 and the bitmap data output from the frequency-specific banding unit 106 are input, and the halftone data (HT) or the bitmap data (B) is input according to a function selection signal given from the CPU.
T) is selectively output. Halftone processing unit 4
Numeral 02 is provided in parallel with the copy-type compression process, receives multi-valued image data having continuous tone, and outputs halftone data (HT) obtained by halftone processing the multi-valued image data to the selector 401. I do. The restoring process is the same as in the first embodiment.

In the above image processing apparatus, the function selection signal for selecting the facsimile system is sent from the CPU to the selector 40.
When input to 1, halftone data output from the halftone processing unit 402 is selected by the selector 401 and input to the JBIG encoding unit 108. And JBIG
The encoded data output from the encoding unit 108 is read from the image storage memory 109 and transmitted by facsimile. A function selection signal for selecting a copy system is supplied to a selector 401.
, The compression / accumulation / decompression processing is executed in the same manner as in the first embodiment.

Note that facsimile transmission can be performed even if a simple binarization processing unit is provided in place of the halftone processing unit 402 in the fourth embodiment.

FIG. 14 is an overall configuration diagram of a multifunction peripheral provided with any one of the image processing apparatuses according to the first, second, third, and fourth embodiments. In the figure, reference numeral 500 denotes the first embodiment.
The scanner 501 is one of the image processing apparatuses 2, 3, and 4.
The printer 502 is a multi-level printer that prints multi-level image data compressed, stored, and restored by the image processing apparatus 500. When the JBIG encoding unit 108 of the image processing apparatus 500 is used as an encoder as in the fourth embodiment, the compressed data stored in the image storage memory 109 is transmitted from the modem 503 to a public line. The image processing apparatus 500 receives rotation control data and a function selection signal from the CPU 504 via an internal bus.

It should be noted that the combination of bitmap data subjected to banding processing by frequency and JBIG encoding as described above can achieve extremely high compression efficiency.
Instead of encoding, other encoding, for example, MH, MR, etc. can be used.

Further, the image processing apparatus of the first embodiment may be appropriately combined with the second, third, and fourth embodiments to form an image processing apparatus and a multifunction peripheral.

[0075]

As described in detail above, according to the present invention, even if edited after image compression, restoration can be performed accurately and without image quality degradation, the required increase in memory capacity can be suppressed, and JBIG or the like can be suppressed. An image processing apparatus that realizes image encoding / decoding that is compatible with encoding / decoding according to the facsimile standard can be provided.

[Brief description of the drawings]

FIG. 1 is a block diagram of an image processing apparatus according to a first embodiment;

FIG. 2 is a base pattern diagram used in the Haar transform

FIG. 3 is a diagram showing HAAR transform coefficients rearranged in the order of frequency components.

FIG. 4 is a conceptual diagram of a quantization process in the first embodiment.

FIG. 5 is a diagram showing a bitmap pattern subjected to banding processing by frequency in the first embodiment.

FIG. 6 is a bitmap pattern diagram in which a sample is actually band-processed by frequency.

FIG. 7 is a conceptual diagram of an inverse quantization process according to the first embodiment.

FIG. 8 is a block diagram of an image processing apparatus according to a second embodiment;

FIG. 9 is a conceptual diagram of a block rotation editing process.

FIG. 10 is a conceptual diagram of an in-block rotation process.

FIG. 11 is a conceptual diagram of an image synthesis / editing process.

FIG. 12 is a block diagram of an image processing apparatus according to a third embodiment;

FIG. 13 is a block diagram of an image processing apparatus according to a fourth embodiment;

FIG. 14 is an overall configuration diagram of a multifunction peripheral.

FIG. 15 is a block diagram of a conventional example in which editing processing is performed before JPEG encoding.

FIG. 16 is a block diagram of a conventional example in which JBIG encoding is performed after performing halftone processing;

[Explanation of symbols]

 Reference Signs List 101 first line memory 102 block dividing unit 103 Haar transform unit 104 quantizing unit 105 block data generating unit 106 band-by-frequency unit 107 second line memory 108 JBIG encoding unit 109 image storage memory 110 JBIG decoding unit 111 Third line memory 112 Block data restoration unit 113 Frequency component restoration unit 114 Inverse quantization unit 115 Inverse Haar transform unit 116 Fourth line memory 117 Block synthesis unit 201 Block rotation edit processing unit 202 Page memory 203 Block synthesis rotation unit 301 Rotation information header adding unit 302 Rotation information header detection unit 401 Selector 402 Halftone processing unit

──────────────────────────────────────────────────続 き Continued on the front page (58) Field surveyed (Int.Cl. ⁷ , DB name) H04N 1/41-1/419 H04N 1/387

Claims (11)

(57) [Claims] 1. An orthogonal transformation means for decomposing multi-valued image data of a predetermined block size into orthogonal transformation coefficients corresponding to spatial frequencies, and a quantization means for transforming the orthogonal transformation coefficients into a predetermined number of bits for each spatial frequency. Banding means for rearranging the quantized data into a banded arrangement for each spatial frequency to generate bit serial data in which quantized data of the same spatial frequency band continues between adjacent blocks, and compresses the bit serial data. An image processing apparatus comprising: an encoding unit. 2. An image processing apparatus according to claim 1, further comprising an editing unit configured to perform image editing on a block-by-block basis with respect to the quantized data calculated by the quantization unit. The image processing apparatus according to claim 1, wherein the arrangement is rearranged into an arrangement. 3. Multivalue image data of a predetermined block size is
Orthogonal transform that decomposes into orthogonal transform coefficients corresponding to spatial frequencies
Means for converting the orthogonal transform coefficients into predetermined bits for each spatial frequency.
Quantization means for converting the quantized data into spatial frequency
Sorted into a banded arrangement for each number and between adjacent blocks
Consecutive bits of quantized data in the same spatial frequency band
Banding means for generating serial data;
JBIG code that compresses serial data using the JBIG method
An image processing apparatus comprising: 4. A decoding means for expanding data compressed by the coding means, a frequency component restoring means for restoring quantized data of each spatial frequency component for each block from the decoded bit serial data, Inverse quantization means for inversely quantizing the restored quantized data of each spatial frequency component to restore orthogonal transform coefficients; inverse orthogonal transform means for restoring the original block image from the restored orthogonal transform coefficients; claim from claim 1 to block image synthesis to a comprises a block combining means for generating the original multi-valued image data, was
3. The image processing device according to any one of 3 . 5. The image processing apparatus according to claim 1, wherein the editing unit controls a write address or a read address of a memory to which the quantized data is written in accordance with rotation control data indicating a rotation amount and a rotation direction of the image, and rotates the image in block units. 5. The block synthesizing unit according to claim 4 , wherein the block synthesizing unit controls a write address or a read address of a memory in which the restored block image is written according to the rotation control data to rotate the image in each block. Image processing device. 6. An information adding means for adding rotation information indicating the contents of rotation control to the bit serial data of an image rotated by said editing means in page units, and said rotation information is obtained from decoded bit serial data. And information detecting means for detecting
Claims, characterized in that for controlling the rotation of the block in the image according to the rotation information detected by said information detection means
Item 6. The image processing device according to Item 5 . 7. The facsimile coding / decoding method according to claim 4, wherein the coding method of said coding means and the decoding method of said decoding means are facsimile coding / decoding methods. Image processing device. 8. A halftone processing means for converting the multi-valued image data into halftone data, and switching between the halftone data or the bit serial data in accordance with a function selection signal for instructing a copy operation or a facsimile transmission. ; and a function selecting means for inputting to said coding means, 請 If facsimile transmission is selected the coded data, from claim 4, characterized in that the external output
The image processing device according to claim 7 . 9. An image input means for scanning a document and inputting image data, an image processing apparatus according to claim 4, and recording an image restored by said image processing apparatus. A multifunction peripheral comprising: a recording unit; and a communication unit that performs facsimile transmission of an image for which facsimile transmission has been instructed among images compressed by the image processing apparatus. 10. The multi-valued image data cut out to a predetermined block size is decomposed into orthogonal transform coefficients corresponding to spatial frequencies, and the orthogonal transform coefficients are converted into a predetermined number of bits for each spatial frequency. Encoding, wherein the data is rearranged into a banded arrangement for each spatial frequency to generate bit serial data in which quantized data of the same spatial frequency band continues between adjacent blocks, and the bit serial data is compressed. Method. 11. Decomposing multi-valued image data cut into a predetermined block size into orthogonal transform coefficients corresponding to spatial frequencies, converting the orthogonal transform coefficients into a predetermined number of bits for each spatial frequency, The data is rearranged into a banded arrangement for each spatial frequency to generate bit serial data in which quantized data of the same spatial frequency band continues between adjacent blocks, and the bit serial data is compressed and stored while compressed. Read and decrypt the data
The quantized data of each spatial frequency component is restored for each block from the decoded bit serial data, and the quantized data of each restored spatial frequency component is dequantized to restore the orthogonal transform coefficient. An image processing method comprising restoring an original block image from transform coefficients, and combining the restored block image to generate original multi-valued image data.